- Prompt Gamma Neutron Activation Analysis
Papers on the nonlinear problem of determining composition of large samples using neutron induced prompt gammas
- Time Dependent Radiation Transport
We have been developing methods to solve the time-dependent radiation transport equation using finite volume methods and Riemann solvers in one and two dimensions.
- Complex System Monitoring
Research on monitoring complex systems such as nuclear plants.
- Determination of Reactivity for MUSE-4 SC0 Configuration
The MUSE-4 experiment was designed to study the space-time characteristics of accelerator driven fission systems. This paper from the Transactions of the American Nuclear Society, Volume 90, describes a reactivity monitor suitable for such systems.
- Probabilistic Techniques for Diagnosis of Multiple Component Degradations
We present probabilistic techniques that make synergistic use of available process information
for diagnosis and detection of component fault manifestation in a multicomponent system. We begin by
describing the motivation for using probabilistic techniques for systems diagnostics and then define
probabilistic expressions that embody the diagnostics knowledge of interest. We show that a combination
of a Bayesian expression with the solution to the Chapman-Kolmogoroff equation contains the diagnostic
information of interest while explicitly making use of available process information including plant data
or measurements, mathematical system models, and individual component reliability data. Given these
probabilistic expressions, we introduce a practical means of obtaining the necessary constituent probability
density functions corresponding to feasible component transitions via an adaptive Kalman filtering
formulation. To demonstrate the consolidated probabilistic technique, we consider a low-order model of a
balance of plant of a boiling water reactor, represented by 11 system variables, 9 component characteristics,
and 5 observations. We simulate 5 to 10% degradations in two components subject to 1% signal
noise in two different transient events. Our test calculations indicate that the proposed algorithm is able to
provide correct fault detection and diagnosis of the faulted components and fault magnitudes, together
with a rank-ordered likelihood of the binary faults.